Roderic A. Girle

Decision Support for Practical Reasoning

Practical reasoning is reasoning about what is to be done. A decision on what to do may involve weighing the options open to an individual, taking into account dependencies on the actions of others, or complex collaborative decision-making. The role of argument in practical reasoning is explored in this chapter, both from a philosophical and computational perspective. In doing so, we discuss the use of computational systems in assisting people engaged in decision making, and, in particular, we investigate practical reasoning as joint deliberation between the human and decision support system. Such a system, it is argued, facilitates research into the use of argumentation techniques in computational models of practical reasoning, and the use of computational models to evaluate theories of practical reasoning.

Argumentation in Multi-Agent Systems, Fifth International Workshop, ArgMAS 2008

Argument-Based Reasoning.- Personality-Based Practical Reasoning.- Argumentation Based Resolution of Conflicts between Desires and Normative Goals.- A Constrained Argumentation System for Practical Reasoning.- An Argumentation Framework Based on Strength for Ontology Mapping.- Contextual Extension with Concept Maps in the Argument Interchange Format.- Argumentation and Dialogue.- Command Dialogues.- Argumentation and Artifact for Dialogue Support.- Co-ordination and Co-operation in Agent Systems: Social Laws and Argumentation.- Annotation and Matching of First-Class Agent Interaction Protocols.- Strategic and Pragmatic Issues.- Argumentation- vs. Proposal-Based Negotiation: An Empirical Case Study on the Basis of Game-Theoretic Solution Concepts.- Argumentation-Based Information Exchange in Prediction Markets.- An Argumentative Approach for Modelling Coalitions Using ATL.- A Dialogue Mechanism for Public Argumentation Using Conversation Policies.

Methodological Issues for the Logic of Questions and Commands

There has been much recent interest in logics for questions and commands. The authors approve, but they argue that methodological issues must be addressed, before it is possible to understand what such logics are for and what they should be like. In particular, the authors deny that the formulas in such logics correspond directly to sentences in ordinary language. Logic is not linguistics. What then are the semantics for the formulas of logics of questions and commands? The focus here is mostly on questions. The authors argue that logics designed to capture the conditions for correct reasoning involving questions require a semantics that treats question-answer pairs as values. They also argue that formal dialogue approaches to the logic of questions should be interpreted in the light of the denial that logic is about language.

Logical Fiction: Real vs. Ideal

Formal systems for knowledge and belief, from Lemmon13 and Hintikka11 to present day Belief Revision systems5, have often been described as models of “ideal rational agents.” From the first, there has been extensive controversy about the validity of the models. 12, 14, 18 A series of topics have given focus to the controversy. They include distinguishing knowing from believing, formalising positive and negative introspection, omniscience of various kinds, the contrast between finite and infinite, and contradictory belief. We consider the extent to which a range of formal models of knowledge and belief are reasonable and realistic. We conclude with comments on the persistence of unreal models and the lack of discussion of their structure.

Shades of consciousness

It has been argued that consciousness might be what differentiates human from machine mentality. What then is consciousness? We discuss consciousness, particularly perception accounts of consciousness. It is argued that perception and consciousness are distinct. Armstrongs (1980) account of consciousness is rejected. It is proposed that perception is a necessary but not sufficient condition for consciousness, and that there is a distinction to be drawn between consciousness and self-consciousness. Consciousness is tightly linked to attention and to certain sorts of knowledge. Implications for machine consciousness and machine attention are discussed.

Dialogue logic as dynamic logic

There are several formal systems for persuasive dialogue. Dialogue systems are multi-agent systems, and this contrasts with the general lack of any agency in standard logics other than in the case of epistemic and deontic logics. Dialogue systems have been called logics. A logic usually has a semantics and a proof sys- tem, and questions of soundness and completeness arise. Any dialogue conducted according to the rules of a dialogue logic is a complex process. Dynamic Logic is a logic of processes, with a possible world semantics. This paper is a preliminary exploration of transforming the dialogues of dialogue logic into complex Dynamic Logic processes with emphasis on semantic models. The transformation gives rise to many questions, three of which are discussed. Dialogue logic includes commit- ment sets which behave in a way similar to belief sets and undergo changes during any dialogue. There are issues as to the relationship between belief and commitment, and whether the logics of belief change apply to commitment stores. There is also the issue of the logical nature of the evaluation of dialogues as legal or illegal.

The question of the question in critical thinking

The question of the question in Critical Thinking is the same as in logic at large. The basis on which Critical Thinking is taught is usually the logic of deductive premise-conclusion arguments. The basic elements of the logic are propositional. This approach gives no basis for teaching students to think logically about either questions or commands. At the same time questions and commands are an important part of practical life. In either asking or responding to questions it is easy to commit any of several illogical moves. Although the same applies to commands, this paper will focus on questions. We raise the question of coherent systems in which questions occupy a proper place, systems which could form an appropriate theoretical background for teaching Critical Thinking and propose dialogue logic.

Exploring the epistemic labyrinth: new directions in the formal theory of knowledge representation

Recently enormous interest has been generated in what researchers in AI and Computer Science have termed the formal theory of knowledge representation or what is known to logicians as epistemic logic. New life has been breathed into this field as the means and motivation for the computational investigation of these formal theories is now at hand. Much of this work has arisen from the epistemic interpretation of normal modal logics. However as vehicles for knowledge representation it now seems clear that these systems are completely inadequate. In this paper we investigate the alternative posed by the non-normal modal logics and discuss techniques for constructing efficient automated theorem provers for these systems.